Designer Receptor Exclusively Activated by Designer Drug (DREADD)-Mediated Activation of the Periaqueductal Gray Restores Nociceptive Descending Inhibition After Traumatic Brain Injury in Rats

Abstract

Traumatic brain injury (TBI) patients frequently experience chronic pain that can enhance their suffering and significantly impair rehabilitative efforts. Clinical studies suggest that damage to the periaqueductal gray matter (PAG) following TBI, a principal center involved in endogenous pain control, may underlie the development of chronic pain. We hypothesized that TBI would diminish the usual pain control functions of the PAG, but that directly stimulating this center using a chemogenetic approach would restore descending pain modulation. We used a well-characterized lateral fluid percussion model (1.3 ± 0.1 atm) of TBI in male rats (n = 271) and measured hindpaw mechanical nociceptive withdrawal thresholds using von Frey filaments. To investigate the role of the PAG in pain both before and after TBI, we activated the neurons of the PAG using a Designer Receptor Exclusively Activated by Designer Drug (DREADD) viral construct. Immunohistochemical analysis of brain tissue was used to assess the location and confirm the appropriate expression of the viral constructs in the PAG. Activation of the PAG DREADD using clozapine N-oxide (CNO) caused hindpaw analgesia that could be blocked using opioid receptor antagonist, naloxone, in uninjured but not TBI rats. Due to the importance of descending serotonergic signaling in modulating nociception, we ablated spinal serotonin signaling using 5,7-DHT. This treatment strongly reduced CNO-mediated anti-nociceptive effects in TBI but not uninjured rats. To define the serotonergic receptor(s) required for the CNO-stimulated effects in TBI rats, we administered 5-HT₇ (SB-269970) and 5-HT_1A (WAY-100635) receptor antagonists but observed no effects. The selective 5-HT_2A receptor antagonist ketanserin, however, blocked CNO's effects in the DREADD expressing TBI but not DREADD expressing sham TBI animals. Blockade of alpha-1 adrenergic receptors with prazosin also had no effect after TBI. Descending pain control originating in the PAG is mediated through opioid receptors in uninjured rats. TBI, however, fundamentally alters the descending nociceptive control circuitry such that serotonergic influences predominate, and those are mediated by the 5-HT_2A receptor. These results provide further evidence that the PAG is a key target for anti-nociception after TBI.

Introduction

Traumatic brain injury (TBI) is a sudden injury to the brain caused by an external force, such as a violent blow, pressure wave, or penetrating injury to the head. TBI is considered to be a leading cause of death and disability worldwide, with an estimated 500-800 new cases of TBI per 100,000 people each year in the U.S. alone.¹ Survivors of TBI may live with several adverse outcomes including impairments in cognition, sensory processing, motor control, and chronic pain that can all have a negative impact on the life of the patient.^2–7 A systematic review of human patients concluded that mild TBI was associated with a 75.3% rate of pain.⁸ The most common type of pain after TBI is post-traumatic headache (PTH),⁹ with about 90% of patients with a history of mild TBI showing features of PTH. Other frequently reported painful areas include the back and upper limbs, with the overall pain prevalence ranging from 22 to 95%, depending on the site. Therefore, identifying the mechanisms through which chronic pain develops after TBI and generating potential treatments is of vital importance to the field of neurotrauma.

Endogenous modulation of pain signaling is controlled by a balance of descending inhibitory and facilitatory pathways connecting the brainstem to the spinal cord. We and others have shown that TBI causes damage to endogenous pain control centers in the brainstem, tipping the balance of inhibition versus facilitation of descending pain modulation in favor of facilitation.^6,7,10–13 The periaqueductal gray (PAG) is a major regulator of these descending pathways, which indirectly communicates with the spinal cord largely via the rostral ventromedial medulla (RVM).^14-16 Fibers from the RVM that innervate the spinal dorsal horn are predominantly serotonergic, the effect of which can be either inhibitory or facilitatory, depending on the receptor subtype activated.^17–20 The PAG also has a prominent role in opioid antinociception. Administration of opioids directly into the PAG of rodents,^21-23 as well as electrical stimulation of the PAG in humans produces antinociception.^24,25 This antinociception can be attenuated by concurrent administration of an opioid antagonist such as naloxone, indicating that antinociception evoked from the PAG is opioid receptor-mediated.^26-28

Our studies using experimental models of TBI in rats and mice have shown that changes in nociceptive signaling after TBI manifest in two distinct phases. The initial serotonin-dependent phase occurs within 24 h of TBI and involves mechanical allodynia of the hindlimbs dependent on spinal 5-HT3 receptors that resolves within a month of injury.²⁹ During this phase, there are also significant increases in neuroinflammation within several key areas of the brain and spinal cord that are involved in descending pain modulation.¹⁰ This is followed by a more slowly developing second phase characterized by the failure of a critical endogenous pain control mechanism known as diffuse noxious inhibitory control (DNIC) or conditioned pain modulation (CPM) in humans.¹¹ Recently, the term “descending control of nociception” (DCN) has been adopted to describe “DNIC-like” pathway in conscious behaving animals and will be used herein.^30,31 Although noradrenergic signaling has been shown to play a key role in the DCN response by some,^19,32-34 it was observed that pharmacologically blocking noradrenaline reuptake was insufficient to restore DCN in TBI rats.¹¹ Enhancement of descending serotonergic signaling could restore the DCN response in TBI rats.¹¹

The functional properties of the PAG have been shown to differentiate patients with chronic pain from healthy, uninjured individuals and these differences correlate with their individual CPM response.^35–38 Chronic pain patients typically have a poorer CPM response compared with healthy uninjured patients. This suggests that functional changes of the PAG could be considered chronic pain predictors.^35–38 Clinical evidence has shown that TBI can cause damage to endogenous pain control centers in the brainstem, including the PAG. For example, an magnetic resonance imaging study revealed that the fractional anisotropy (FA; a measure of orderly tissue structure) of the PAG was diminished in TBI patients suffering from chronic pain.³⁹ It has also been shown that decreased functional connectivity of the PAG with key regions of the default mode network (DMN) occurred in TBI patients with acute post-traumatic headache.⁴⁰ Most recently, changes to the connections between the PAG and sensorimotor cortex during the early acute stage after TBI (< 72 h) could predict whether the TBI patient was likely to develop chronic pain.³⁸

We designed a strategy to activate the neurons of the PAG using a Designer Receptor Exclusively Activated by Designer Drug (DREADD) construct. The PAG is a key part of the descending pathway that modulates nociception, fear and anxiety behaviors in both humans and rodents.^41,42 This allowed us to study the anti-nociceptive efficacy and mechanisms of this critical center in both control and TBI conditions.

Methods

This section was assembled using material and methods from our previous publications.^10,11,29,43

Animals

All studies were approved by the Veterans Affairs Palo Alto Health Care System Institutional Animal Care and Use Committee (Palo Alto, CA, USA) and followed the animal guidelines of the National Institutes of Health (NIH) Guide for the Care and Use of Laboratory animals (NIH Publications 8th edition, 2011).⁴⁴ Male Sprague Dawley rats (n = 271, 250-300 g; Envigo (Indianapolis, IN, USA), were housed under standard conditions with a 12 h light–dark cycle (6:30 a.m. to 6:30 p.m.) and were given food and water ad libitum. The animals were housed in pairs in 30 × 30 × 19-cm isolator cages with solid floors covered with a 4-cm layer of wood chip bedding. Experimenters were blinded to the identity of treatments and experimental conditions. All studies were designed to minimize the number of rats required. All in vivo experiments were performed between 6 a.m. and 2 p.m. in our facility's Veterinary Medical Unit.

Drugs

Drugs used included: prostaglandin E2 (PGE-2; 1 μg/50 μL), intraplantar (i.pl.) (14010, Cayman Chemicals, MI, USA), a principal mediator of inflammation and pain hypersensitivity; capsaicin (CAP; 10 μg/10 μL), i.pl. (M2028; Millipore-Sigma, MO, USA), which causes hyperalgesia to mechanical stimuli at and around the injection site; clozapine N-oxide (CNO; 3 mg/kg), intraperitoneally (i.p.; 16882, Cayman Chemicals), a ligand used to selectively activate DREADDs and is otherwise pharmacologically and behaviorally inert in animals at this dose; SB-269970 (SB-269970; 10 μg/10 μL) intrathecally (i.t.; 17081; Cayman Chemicals), a 5-HT₇ receptor antagonist; 5, 7-dihydroxytryptamine (DHT; 60 μg/20 μL), i.t. (CDX-H0026; Adipogen Corp, CA, USA), a neurotoxin that destroys 5-HT axons and nerve terminals when injected into the brain and/or spinal cord; desipramine hydrochloride (30 mg/kg), i.p. (3067; Tocris Bioscience, MN, USA), to prevent the partial depletion of norepinephrine that can also occur following DHT injection; WAY-100635 maleate (WAY; 3 μg/10 μL), i.t. (4380; Tocris), a 5-HT_1A receptor antagonist; ketanserin tartrate (KET; 30 μg/10 μL), i.t. (22058; Cayman Chemicals), a 5-HT_2A receptor antagonist; naloxone hydrochloride dihydrate (NAL; 1 mg/kg), i.p. (N7758; Millipore-Sigma), an opioid receptor antagonist; and prazosin (PRZ; 1 mg/kg), i.p. (15023; Cayman Chemicals), an α-1-AR antagonist. The doses of these drugs were chosen based on previous studies⁴⁵ by our lab and/or peer reviewed papers that have shown that the listed concentrations are adequate to block the 5-HT_1A⁴⁶ or 5-HT_2A⁴⁷ receptor.

Viral vectors

Designer Receptor Exclusively Activated by Designer Drugs or DREADDS are often genetically modified muscarinic acetylcholine receptors that are conjugated to an adeno-associated virus (AAV) and include specific promoters to limit their expression to particular cell populations.^48–51 Depending on the type of DREADD receptor used, they can either excite (hM3Dq) or inhibit (hM4Di) the activity of the transduced neuron.^52,53 The AAV used in the current study, pAAV-hSyn-hM3D(Gq)-mCherry, was a gift from Bryan Roth (Addgene viral prep # 50474-AAV8). pAAV-hSyn-hM3D(Gq)-mCherry contains the human synapsin promoter, hSyn,⁵⁴ to selectively express the excitatory hM3Dq DREADD receptor fused with mCherry⁵² in neurons of the injected area, in this case, the PAG. Activation of the DREADD receptor is accomplished through systemic delivery of the otherwise inert exogenous ligand clozapine-N-oxide (CNO) The control AAV8 vector contained the hSyn promoter with only a green fluorescent protein (GFP) reporter protein without the DREADD gene, hM3Dq, pAAV8-hSyn-eGFP.

DREADD microinjection surgeries

All surgeries were performed under isoflurane anesthesia (5% for induction and 2% for maintenance). Rats were secured in a stereotactic frame, and the coordinates for the left and right PAG (from Bregma: -7.8 mm A-P, ±0.5 mm M-L, 5.8 mm D-V) were entered and marked on the skull with a surgical pen.⁵⁵ A 3-mm craniotomy was made to accommodate both marked locations using a mini-drill, a 2.1 mm and 1.4 mm drill burr (Fine Science Tools, CA, USA); 0.3 μL of the excitatory DREADD (AAV8-hSyn-xhM3DqmCherry) or the control virus (AAV8-hSyn-eGFP) was injected at a rate of 0.03 μL/min using a 2-inch 30G beveled needle (7803-07, Hamilton, NV, USA) connected to a Hamilton 5 μL syringe (7634-01, Hamilton). The needle was left in place for 10 min after infusion and then slowly removed. The overlying scalp wound was closed using staples and rats were allowed 4 weeks for recovery and viral expression.

TBI surgery

The lateral fluid percussion (LFP) rat model of TBI was used as described previously.^10,11,29 Briefly, a midline incision was made in the scalp, and a 5-mm craniotomy was made on the right side of the skull using a mini-drill and a 5 mm trephine burr (Fine Science Tools). The craniotomy was placed midway between the bregma and lambda sutures, and 2 mm right of the midline suture. Using cyanoacrylate glue and dental acrylic, a female luer attachment was affixed to the craniotomy opening. Following recovery of pinch reflexes, the luer attachment was connected to the LFP apparatus (Amscien Instruments, VA, USA), and a pressure wave of 1.3 (± 0.1 atm) to produce mild level injuries or no pressure wave (uninjured) was applied to rat dura based on previous reports. Thereafter, the luer attachment and dental acrylic were removed, the bone flap replaced, and the overlying wound closed using staples. Rats were allowed to recover in their home cages.

Intrathecal injection

Rats were anesthetized with isoflurane throughout the procedure and the paw pinch reflex was used to ensure the state of anesthesia. A 3 cm² window of fur, near the base of the tail, was shaved and washed with 70% ethanol to maximize visualization during needle insertion. An empty 50 mL conical centrifuge tube was placed underneath the rat to arch the lumbar vertebral column, and the spinous process of L5 was located. The spinous process of L5 was pulled in a cranial direction and the vertebral body of L6 was pulled caudally towards the tail. This maximized the space within the groove between L5 and L6 vertebrae into which a 25 G needle was carefully inserted. Successful entry of the needle into the intradural space was confirmed by the observation of a tail flick. The needle was immediately secured into position with one hand and the desired volume of substance slowly injected with the other hand. Once injection was performed, the rat was placed on a warming pad prior to returning to their home cage.

For spinal 5-HT depletion, the serotoninergic neurons of the lumbar spinal cord were targeted using an intrathecal injection of 5, 7-dihydroxytryptamine diluted in sterile saline (DHT, 60 μg/20 μL, i.t., CDX-H0026; Adipogen Corp, CA, USA). DHT is a neurotoxin that destroys 5-HT axons and nerve terminals when injected into the brain and/or spinal cord. To prevent the partial depletion of norepinephrine that can also occur following DHT injection, all rats were pretreated with desipramine hydrochloride diluted in sterile saline (30 mg/kg, i.p., 3067; Tocris Bioscience, MN, USA) 45 min prior to DHT administration. Any behavioral experiments began 3 days post-injection to allow the DHT to have maximal toxic effect.

Behavioral testing

Mechanical withdrawal thresholds were measured using a modification of the up-down method and von Frey filaments as described previously.⁵⁶ Animals were placed in clear Plexiglas boxes (17 cm length × 11 cm width × 13 cm height) on an elevated horizontal wire mesh stand (IITC Life Science Inc, CA, USA). After 60 min of acclimation, fibers of increasing stiffness with initial bending force of 4.31 N (ranging from 4.31 N up to 5.18 N) were applied to the plantar surface of the hindpaw, and left in place for 5 sec with enough force to slightly bend the fiber. Withdrawal of the hindpaw from the fiber was scored as a response. When no response was obtained, the next firmer fiber in the series was used in the same manner. If a response was observed, the next less stiff fiber was applied. Testing continued until four fibers had been applied after the first one causing a withdrawal response, allowing the estimation of the mechanical withdrawal threshold using a curve fitting algorithm.⁵⁷

DCN paradigm

Assessment of post-traumatic mechanical hypersensitivity was done using a variant of a widely used noxious stimulation–induced anti-nociceptive protocol.^58-60 Briefly, the rats were confirmed to recover to baseline mechanical thresholds after TBI (49 days post-injury). Next, intraplantar injection of PGE-2 (test stimulus) into the hindpaw contralateral to the TBI was used to produce brief hypersensitivity.⁶¹ After an hour, the PGE-2 injection hindlimb withdrawal thresholds were measured using the von Frey filaments as described above. Subsequent to mechanical threshold evaluations, rats are typically given a capsaicin injection (conditioning stimulus) into the dorsal surface of the forepaw ipsilateral to the TBI to induce DCN. This protocol was used for experiments detailed in Figure 7. In contrast, for all other experiments that involved DREADD-injected rats, capsaicin was not used. Instead, an i.p. injection of CNO resulting in the activation the PAG was used to induce an antinociceptive response. All intraplantar and i.p. injections were administered under light isoflurane anesthesia. At 15, 30, 60, 120, 180, and 240 min after capsaicin or CNO injection, withdrawal thresholds were measured using the von Frey filaments as described above.

FIG. 7.

The effect of 5-HT₇ receptor antagonist, SB-269970 on the temporary reinstatement of diffuse noxious inhibitory control (DNIC) following pretreatment with the selective serotonin reuptake inhibitor, escitalopram (ESC) at 49 days post-traumatic brain injury (TBI). (A) Both sham and TBI rats were given a daily injection of ESC intraperitoneally (i.p.; 10 mg/kg) for 3 days prior to DNIC testing. On the day of testing, half the sham and TBI were randomly selected and given an intrathecal (i.t.) injection of SB-26 following the test stimulus (prostaglandin E2 [PGE-2]). (B) Behavioral assessment on day 49 post-injury revealed that SB-269970 treatment blocked the ESC mediated restoration of DNIC in TBI rats. Neither SB-269970 nor ESC treatment had any effect on DNIC processing in the sham-injured rats. White arrow - PGE-2; intraplantar (i.pl.) [1 μg/50 μL]. Black arrow - Capsaicin; i.pl. [10 μg/10 μL]. Three-way mixed repeated measures analysis of variance revealed a significant main effect of injury condition [F(1,20) = 117.60, p = 7.96 × 10⁻¹⁰, ƞ2 = 0.86], SB-269970 [F(1,20) = 67.23, p = 7.96 × 10⁻⁸, ƞ2 = 0.77], time [F(6,120) = 38.15, p = 1.28 × 10⁻²⁵, ƞ2 = 0.66], injury by SB-269970 interaction [F(1,20) = 21.55, p = 1.5 × 10⁻⁴, ƞ2 = 0.52], time by injury interaction [F(6,120) = 3.93, p = 0.001, ƞ2 = 0.16], time by SB-269970 interaction [F(6,120) = 5.15, p = 9.7 × 10⁻⁵, ƞ2 = 0.21], time by injury by SB-269970 [F(6,120) = 4.61, p = 3.0 × 10⁻⁴, ƞ2 = 0.19]. Post hoc analysis reveals that the restorative effect of escitalopram pretreatment is reversed by SB-269970 treatment (7B). *TBI/ESC/SB-269970 vs. TBI/ESC/VEH (p < 0.001) by Bonferroni's post hocs. Data are presented as mean ± standard error of the mean (n = 8 to 12).

Immunohistochemistry

This protocol has been described in detail previously.^10,29 Briefly, rats were perfused with 4% paraformaldehyde, brains were then removed and cryoprotected in 20% sucrose in phosphate buffered saline (PBS) for 2 days at 4°C. The brainstems were cut into 12-μm transverse sections using a cryostat. Immunostaining was performed using an antibody against mCherry (1:500, MA5 32977, ThermoFisher, MA, USA) or GFP (1:400, A1120; ThermoFisher, USA). Blocking of all the sections took place at room temperature for 2 h in PBS containing 10% normal goat serum (Vector Laboratories, CA, USA), followed by exposure to the primary antibody overnight at 4°C. mCherry and GFP were visualized with Alexa Fluor-conjugated second antibodies (ThermoFisher, MA, USA). Controls prepared with primary antibody omitted showed minimal background fluorescence under the conditions employed. Staining was performed concurrently for each group of sections compared with one another, and photographed under identical conditions.

Quantitative real-time polymerase chain reaction

Total RNA from rat spinal cord was extracted by using the RNeasy Mini Kit (Qiagen, CA, USA), and the purity and concentration were determined spectrophotometrically. Then complementary DNA was synthesized from 1 μg RNA using RT² First Strand Kit (Qiagen, Hilden, Germany). Real-time polymerase chain reactions were performed on an ABI 7900HT sequencing detection system (Applied Biosystems, Waltham, MA) using RT² qPCR Primer Assay and the RT² SYBR Green qPCR Mastermix (Qiagen, Hilden, Germany). 5-HT_2A and 18S primer sets were validated on dissociation curves to document single product formation. The data from real-time polymerase chain reaction experiments were analyzed as described in the manufacturer's manual for the ABI 7900HT sequencing detection systems.

Image analysis

In all data assessments, both the photography and image analysis were performed by observers who were blinded to the experimental conditions. mCherry (DREADD virus) or GFP (control virus) expression was evaluated bilaterally in the periaqueductal gray (PAG, A-P -7.0 to -8.7 mm) on four coronal sections per brain spaced 120 μm apart. Sections from uninjured, untreated rats were used to establish a threshold level that excluded all non-specific staining. This threshold level was then applied to all experimental groups. Rats with mCherry-positive neurons within one or both sides of the ventrolateral PAG were included in this study. Rats with any mCherry-positive neurons outside of the PAG were excluded. No animals were excluded from this study based on these criteria.

Statistical analysis

An independent statistician (A.R.F.) tested statistical assumptions, performed missing values analysis, descriptive statistical tests, followed by inferential tests. Data met normality, homogeneity of variance, and independence assumptions sufficiently for application of general linear models (GLM). Group comparisons were therefore performed according to a priori balanced experimental designs as two-way, three-way, or four-way mixed analyses of variance (ANOVAs), modeling between- and within-subject factors as appropriate using GLM. Significant effects were followed by interaction plots and Bonferroni's post hocs. All statistical analyses were performed using IBM SPSS software (v.27, IBM Corp.) with base, regression, missing values, and advanced statistics modules. Statistical code is freely available upon request. Graphs were created using Prism 9.1.0 (GraphPad Software). Data are presented as mean values ± standard error of the mean (SEM). All experimental sample sizes (n) were selected by a priori power calculations based on historical data from our laboratory. The figure legends report precise F values, degrees of freedom, p values, effect sizes, and observed power for all significant effects as well as the n's completed for each group.

Data availability

All the raw data are available through the VA- and NIH-supported open data commons for TBI (odc-tbi.org) at https://doi.org/10.3494/F58P4G in support of funder mandates for FAIR (findable, accessible, interoperable, and reusable) data stewardship policies.

Results

DREADD-mediated activation of the Periaqueductal Gray (PAG) can trigger endogenous pain modulation in naïve rats

We first set out to determine whether CNO/DREADD-mediated activation of the PAG was able to trigger an effective analgesic response following PGE-2 sensitization of the hindpaw. For this study, we bilaterally injected rats with the DREADD (AAV8-hSyn-xhM3DqmCherry) or the control (AAV8-hSyn-eGFP) virus into the PAG, 28 days prior to behavioral testing (Fig. 1Ci). Expression of the DREADD-virus (mCherry) was limited to the neurons of the PAG (Fig. 1Cii and Ciii). DREADD expression in these neurons was restricted to the plasma membrane of the neuronal cell bodies and axons in all animals (Fig. 1Civ). Data in Figure 1B reveals that the CNO/DREADD-mediated activation of the PAG did result in an anti-nociceptive response. In contrast, no such response was present in Vehicle (VEH)/DREADD-injected rats and in both CNO/control-injected and VEH/control-injected rats. A three-way mixed repeated measures ANOVA revealed significant main effects of virus condition (p = 6.7 × 10⁻⁷); Drug/CNO (p = 1.0 × 10⁻⁶); time (p = 3.9 × 10⁻¹²). There were significant interactions of time by virus condition (p = 1.45 × 10⁻¹⁰); time by drug interaction (p = 1.11 × 10⁻¹³); and time by drug by virus (p = 3.02 × 10⁻¹²). Post hoc analysis indicated a statistically significant increase in paw withdrawal threshold in DREADD-injected, TBI rats in the presence of CNO at 10, 30, 60, 90, 120, 150, and 180 min compared with vehicle-treated DREADD-expressing TBI rats (p < 0.0001).

FIG. 1.

Designer Receptor Exclusively Activated by Designer Drug (DREADD)-mediated activation of the periaqueductal gray matter (PAG) can trigger endogenous pain modulation in naïve rats. (A) Protocol: Rats were injected with either the DREADD virus (AAV8-hSyn-xhM3DqmCherry) or the control virus (AAV8-hSyn-eGFP) bilaterally into the ventrolateral PAG (B.i.). On the day of testing, 28 days post-injection, all rats were given an intra-plantar injection into the left hindpaw with the test stimulus, prostaglandin E2 (PGE-2; open arrow, intraplantar (i.pl.) [1 μg/50 μL]). An hour after PGE-2 injection, rats were then split into vehicle-treated (saline, intraperitoneally [i.p.]) or clozapine N-oxide (CNO)-treated groups. CNO is used as a ligand to activate DREADD receptors. Low magnification of mcherry positive cells (B.ii and B.iii), High magnification of mcherry positive cells in the periaqueductal gray (B.iii). Positive cells are labeled with a white arrow head (B.iv). (C) Behavioral assessment revealed that CNO-mediated activation of the DREADD virus in the PAG triggered an endogenous pain response that was not present in non-CNO treated, DREADD-injected rats. Further, no change in the Paw withdrawal threshold (PWT) was observed in rats injected with the control virus with or without CNO treatment. Three-way mixed repeated measures analysis of variance revealed significant main effects of virus condition [F(1,17) = 58.46, p = 6.7 × 10⁻⁷, ƞ2 = 0.78]; CNO [F(1,17) = 54.65, p = 1.0 × 10⁻⁶, ƞ2 = 0.76]; time [F(5,85) = 18.00, p = 3.9 × 10⁻¹², ƞ2 = 0.51]; time by virus condition interaction [F(5,85) = 15.05, p = 1.45 × 10⁻¹⁰, ƞ2 = 0.47]; time by CNO interaction [F(5,85) = 21.15, p = 1.11 × 10⁻¹³, ƞ2 = 0.55]; time by CNO by virus condition interaction [F(5,85) = 18.22, p = 3.02 × 10⁻¹², ƞ2 = 0.52]. Post hoc analysis reveals that DREADD/CNO was significantly differed from the other groups over time (1C). White arrow—PGE-2; i.pl. [1 μg/50 μL]. Gray arrow—Clozapine-N-oxide, i.p. [3 mg/kg]). Scale Bar: 100 μm (B.ii) and 50 μm (B.iii and B.iv). Data are presented as mean ± standard error of the mean (n = 5-6, all groups). *DREADD/CNO vs. DREADD/VEH (p < 0.001) by Bonferroni's post hocs.

Intra-PAG injection of DREADD viral vector does not alter nociceptive changes after TBI

It was necessary for the interpretation of our planned experiments using the DREADD construct to confirm that surgery and/or the presence of viral vectors in the PAG did not affect the early nociceptive sensitization characteristic of this TBI model.¹¹ Data in Figure 2 revealed that neither the surgery nor the intra-PAG injection of a virus (DREADD or control) impacted the nociceptive outcomes of TBI over the following 49 days. Two-way mixed repeated measures ANOVA revealed significant main effect of experimental group (p = 3.55 × 10⁻²¹); main effect of time (p = 35.02 × 10⁻¹⁰⁵); group by time interaction (p = 5.11 × 10⁻⁵⁴). Post hoc analysis reveals that all TBI groups were significantly different from the uninjured group (Fig. 2B). Data are presented as mean ± standard error of the mean (SEM; n = 8, all groups).

FIG. 2.

Traumatic brain injury (TBI)-induced nociceptive sensitization of the contralateral hindpaw was unaffected by the injections of the Designer Receptor Exclusively Activated by Designer Drug (DREADD) or control virus into the periaqueductal gray matter (PAG; A) up to 49 days post-injury (B). Two-way mixed repeated measures analysis of variance revealed significant main effect of experimental group [F(3,36) = 166.97, p = 3.55 × 10⁻²¹, ƞ2 = 0.93]; main effect of time [F(7,252) = 227.31, p = 35.02 × 10⁻¹⁰⁵, ƞ2 = 0.86]; group by time interaction [F(21,252) = 28.02, p = 5.11 × 10⁻⁵⁴, ƞ2 = 0.70]. Post hoc analysis reveals that all TBI groups were significantly different from the uninjured group (Fig. 2B). Data are presented as mean ± standard error of the mean (n = 8, all groups).

DREADD-mediated activation of the PAG reduces mechanical sensitization of the hindpaw during the initial pain phase after TBI

We next determined if CNO/DREADD-mediated activation of the PAG could affect the manifestation of the initial nociceptive response to TBI. Figure 3 reveals that CNO/DREADD-mediated activation of the PAG significantly decreased the degree of mechanical sensitization in the hindlimb at both 7 days post-injury (DPI; Fig. 3A) and 21 DPI (Fig. 3B). This effect persisted through the length of the study (approximately 240 min). Analysis 24 h (1440 min) later confirmed that hindlimb hypersensitivity of the CNO/DREADD-injected, TBI rats had returned to pre-CNO levels at both 7 and 21 DPI. Hindpaw sensitivity in VEH/DREADD-injected, TBI rats and in both control-injected, TBI rat groups did not vary from baseline for the duration of the experiment (Fig. 3A, 3B). Three-way mixed repeated measures ANOVA with two nested repeated measures (time and testing day) revealed significant main effect of group (p = 2.43 × 10⁻²⁴); time (p = 8.40 × 10⁻²²); group by time interaction (p = 5.06 × 10⁻⁴¹); No significant main effect of testing day but there was a significant interaction of group by testing day (p = 0.04); no other main effects or interaction were significant. Post hoc analysis reveals that the control/TBI group had significantly lower pain thresholds than the control/uninjured and DREADD treated groups (p < 0.001). However, DREADD/TBI had slightly lower thresholds at 7 DPI compared with 21 DPI. Post hoc examination revealed a decay of the DREADD effect at 240 min for the 7 DPI time-point (Fig. 3A vs. Fig. 3B). Data are presented as mean ± SEM (n = 8, all groups).

FIG. 3.

The effect of Designer Receptor Exclusively Activated by Designer Drug (DREADD)/ clozapine N-oxide (CNO)-mediated activation of the periaqueductal gray matter (PAG) on traumatic brain injury (TBI)-induced nociceptive sensitization at (A) 7 and (B) 21 days post-injury. Assessment of nociceptive sensitivity of the contralateral hindpaw revealed that DREADD-mediated activation of the PAG increases paw withdrawal threshold (peaking 15 min post-CNO) on both 7 and 21 DPI. Gray arrow—CNO, intraperitoneally [i.p.; 3 mg/kg]). three-way mixed repeated measures analysis of variance with two nested repeated measures (time and testing day) revealed significant main effect of group [F(3,28) = 494.16, p = 2.43x10⁻²⁴, ƞ2 = 0.98]; time [F(5,140) = 32.32, p = 8.40 × 10⁻²², ƞ2 = 0.54]; group by time interaction [F(15,140) = 36.73, p = 5.06 × 10⁻⁴¹, ƞ2 = 0.80]; No significant main effect of testing day but there was a significant interaction of group by testing day [F(3,28) = 3.24, p = 0.04, ƞ2 = 0.26]; no other main effects or interaction were significant. Post hoc analysis reveals that the control/TBI group had significantly lower pain thresholds than the control/uninjured and DREADD treated groups. However, DREADD/TBI had slightly lower thresholds at 7 days post-injury (DPI) compared with 21 DPI. Post hoc examination revealed a decay of the DREADD effect at 240 min for the 7 DPI time-point (panels A vs. B). Data are presented as mean ± standard error of the mean (n = 8, all groups). *DREADD/TBI/CNO vs. DREADD/TBI/Vehicle (VEH; p < 0.001).

DREADD-mediated activation of the PAG triggers an analgesic response to PGE-2 in both TBI and uninjured rats

At 49 DPI, after the initial nociceptive phase has resolved, we tested if DREADD-mediated activation of the PAG would trigger an anti-nociceptive response in TBI and uninjured rats to PGE-2 injection in the contralateral hindpaw. The data in Figure 4A demonstrates that when DREADD-injected TBI and uninjured rats were treated with CNO, resulting in PAG activation, it initiated an anti-nociceptive response that significantly reduced hindpaw hypersensitivity which persisted for up to 240 min. In response to the vehicle, the anti-nociceptive response was not observed in DREADD-injected, TBI and uninjured rats (Fig. 4A). Further, CNO treatment did not result in an anti-nociceptive response in control virus-injected, TBI and uninjured rats (Fig. 4B). Four-way mixed repeated measures ANOVA revealed a significant main effect of virus condition (p = 5.66 × 10⁻²⁹); CNO (p = 5.33 × 10⁻³¹); time (p = 1.80 × 10⁻⁴⁰); virus condition by CNO interaction (p = 1.20 × 10⁻²⁹); time by virus condition interaction (p = 1.56 × x10⁻³³); time by CNO interaction (p = 3.12 × 10⁻³⁸); time by virus condition by CNO (p = 3.18 × 10⁻³⁸); no other main effects or interactions reached significance. Post hoc analysis revealed that the DREADD/CNO effects were equally effective in TBI and uninjured subjects indicating reversal of TBI-induced hypersensitivity (Fig. 4B vs. 4C). Data are presented as mean ± SEM (n = 6-8); *DREADD/TBI/CNO versus DREADD/TBI/VEH (p < 0.001) by Bonferroni's post hoc.

FIG. 4.

The effect of the Designer Receptor Exclusively Activated by Designer Drug (DREADD)/ clozapine N-oxide (CNO)-mediated activation of the periaqueductal gray matter (PAG) on the traumatic brain injury (TBI)-induced failure of the diffuse noxious inhibitory control (DNIC) response at 49 days post-injury (DPI) (A). DREADD/CNO-mediated activation of the PAG reinstates DNIC response in TBI rats (B). In absence of CNO-treatment the DNIC response was not restored in DREADD-injected TBI rats (B). Further, DNIC was not restored in the Control-injected TBI rats with or without CNO treatment (C). White arrow—prostaglandin E2 (PGE-2); intraplantar (i.pl.) [1 μg/50 μL]. Gray arrow—Clozapine-N-oxide, intraperitoneally (i.p.; 3 mg/kg). Four-way mixed repeated measures analysis of variance revealed a significant main effect of virus condition [F(1,48) = 612.18, p = 5.66 × 10⁻²⁹, ƞ2 = 0.93]; CNO [F(1,48) = 753.61, p = 5.33x10⁻³¹, ƞ2 = 0.94]; time [F(5,240) = 60.22, p = 1.80 × 10⁻⁴⁰, ƞ2 = 0.56]; virus condition by CNO interaction [F(1,48) = 656.26, p = 1.20 × 10⁻²⁹, ƞ2 = 0.93]; time by virus condition interaction [F(5,240) = 46.59, p = 1.56 × 10⁻³³, ƞ2 = 0.49]; time by CNO interaction [F(5,240) = 55.63, p = 3.12x10⁻³⁸, ƞ2 = 0.54]; time by virus condition by CNO [F(5,240) = 55.61, p = 3.18 × 10⁻³⁸, ƞ2 = 0.54]; No other main effects or interactions reached significance. The results indicate that the DREADD/CNO effectively reversed hypersensitivity. Post hoc analysis revealed that the DREADD/CNO effects were equally effective in TBI and uninjured subjects indicating reversal of TBI-induced hypersensitivity (panels B vs. C). Data are presented as mean ± standard error of the mean (n = 6-8). *DREADD/TBI/CNO vs. DREADD/TBI/Vehicle (VEH; p < 0.001) by Bonferroni's post hocs.

Administration of naloxone (NAL) blocks the CNO/DREADD-mediated analgesic response in uninjured rats but not TBI rats

The PAG is rich in opioid receptors that can modulate nociceptive information and control pain sensitivity. With this in mind, we investigated the role of opioid signaling in the anti-nociceptive response triggered in CNO/DREADD-injected TBI and uninjured rats. Figure 5B reveals that a systemic injection of naloxone (NAL), an opioid receptor antagonist, did block the anti-nociceptive response by the PAG to the PGE-2 injection of the hindpaw in CNO/DREADD-injected uninjured rats. Unexpectedly, NAL failed to block the anti-nociceptive response triggered in CNO/DREADD-injected TBI rats (Fig. 5C). Four-way mixed repeated measures ANOVA revealed a significant main effect of virus condition (p = 4.09 × 10⁻³⁴; naloxone (p = 4.53 × 10⁻⁷); time (p = 9.26 × 10⁻⁷⁴); injury by virus condition interaction (p = 5.0 × 10⁻⁶); injury by naloxone interaction (p = 3.25 × 10⁻¹⁰); virus condition by naloxone interaction (p = 3.39 × 10⁻⁸); injury by virus condition by naloxone (p = 4.70 × 10⁻⁹); time by injury interaction (p = 8.0 × 10⁻⁶); time by virus condition interaction (p = 1.17 × 10⁻⁷²); time by naloxone interaction (p = 1.18 × 10⁻⁸); time by injury by virus condition (p = 7.00 × 10⁻⁶); time by injury by naloxone (p = 2.83 × 10⁻¹²); time by virus condition by naloxone (p = 1.80 × 10⁻¹⁰); time by injury condition by virus condition by naloxone (p = 1.28 × 10⁻¹¹). Post hoc analysis reveals that CNO anti-nociceptive effect is reversed by naloxone in uninjured animals but not in TBI animals (Fig. 5B vs. 5C); DREADD/Uninjured/CNO/NAL vs. DREADD/Uninjured/CNO/VEH (p < 0.001) by Bonferroni's post hocs. Data are presented as mean ± SEM (n = 7-8).

FIG. 5.

The analgesic response to Designer Receptor Exclusively Activated by Designer Drug (DREADD)-mediated activation of the periaqueductal gray matter (PAG) in clozapine N-oxide (CNO)-treated, DREADD-injected, uninjured rats (A) was blocked by the intrathecal treatment with opioid receptor antagonist, Naloxone (NAL) (B). In contrast, naloxone treatment in CNO-treated, DREADD-injected, traumatic brain injury (TBI) rats failed to block the analgesic response to DREADD-mediated activation of the PAG (C). White arrow—prostaglandin E2 (PGE-2); intraplantar (i.pl.) [1 μg/50 μL]. Gray arrow—CNO; intraperitoneally (i.p.; 3 mg/kg). Four-way mixed repeated measures analysis of variance revealed a significant main effect of virus condition [F(1,50) = 942.96, p = 4.09 × 10⁻³⁴, ƞ2 = 0.95]; naloxone [F(1,50) = 33.60, p = 4.53 × 10⁻⁷, ƞ2 = 0.40]; time [F(7,350) = 88.89, p = 9.26 × 10⁻⁷⁴, ƞ2 = 0.64]; injury by virus condition interaction [F(1,50) = 26.37, p = 5.0 × 10⁻⁶, ƞ2 = 0.35]; injury by naloxone interaction [F(1,50) = 61.04, p = 3.25 × 10⁻¹⁰, ƞ2 = 0.55]; virus condition by naloxone interaction [F(1,50) = 42.50, p = 3.39 × 10⁻⁸, ƞ2 = 0.46]; injury by virus condition by naloxone [F(1,50) = 49.95, p = 4.70 × 10⁻⁹, ƞ2 = 0.50]; time by injury interaction [F(7,350) = 5.35, p = 8.0 × 10⁻⁶, ƞ2 = 0.97]; time by virus condition interaction [F(7,350) = 86.88, p = 1.17 × 10⁻⁷², ƞ2 = 0.10]; time by naloxone interaction [F(7,350) = 7.71, p = 1.18 × 10⁻⁸, ƞ2 = 0.13]; time by injury by virus condition [F(7,350) = 5.40, p = 7.00 × 10⁻⁶, ƞ2 = 0.10]; time by injury by naloxone [F(7,350) = 10.74, p = 2.83 × 10⁻¹², ƞ2 = 0.18]; time by virus condition by naloxone [F(7,350) = 9.22, p = 1.80 × x10⁻¹⁰, ƞ2 = 0.16]; time by injury condition by virus condition by naloxone [F(7,350) = 10.19, p = 1.28 × 10⁻¹¹, ƞ2 = 0.17]. Post hoc analysis reveals that CNO anti-nociceptive effect is reversed by naloxone in uninjured animals but not in TBI animals (panels B vs. C). DREADD/Uninjured/CNO/NAL vs. DREADD/Uninjured/CNO/VEH (p < 0.001) by Bonferroni's post hocs. Data are presented as mean ± standard error of the mean (n = 7-8).

Administration of 5, 7-Dihydroxytryptamine (DHT), a selective toxin for serotonergic neurons, blocks the CNO/DREADD-mediated analgesic response by the PAG after TBI

To investigate the potential role of serotonin in the anti-nociceptive response of CNO/DREADD-injected TBI rats, we used 5, 7-Dihydroxytryptamine (DHT), that selectively and persistently ablates spinal serotonergic neurons when injected intrathecally Figure 6A. Data from Figure 6B demonstrates that DHT did block the restoration of the anti-nociceptive response in CNO/DREADD-injected TBI rats compared with VEH treated CNO/DREADD-injected, TBI rats. However, DHT had no effect on the anti-nociceptive response by the CNO/DREADD-injected, uninjured rats. Four-way mixed repeated measures ANOVA revealed a significant main effect of injury (p = 0.007), virus condition (p = 2.92 × 10⁻¹⁸), main effect of DHT (p = 0.025), time (p = 6.32 × 10⁻⁴⁷), injury by virus condition interaction (p = 0.003), injury by DHT (p = 0.035), virus condition by DHT (p = 0.028), injury by virus condition by DHT. Three-way interaction (p = 0.007), time by injury (p = 0.001), time by virus condition (p = 5.92 × 10⁻⁴⁷), time by injury by virus condition (p = 4.2 × 10⁻⁵). Time by injury by DHT (p = 0.043), Time by injury by virus by DHT (p = 0.042). No other main effects or interactions were significant. Post hoc analysis reveals that CNO anti-nociceptive effect is reversed by DHT in TBI animals but not in uninjured animals (p < 0.001; Fig. 6B vs. 6C). Data are presented as mean ± SEM (n = 6-8).

FIG. 6.

The analgesic response to Designer Receptor Exclusively Activated by Designer Drug (DREADD)-mediated activation of the periaqueductal gray matter (PAG) in clozapine N-oxide (CNO)-treated, DREADD-injected, traumatic brain injury (TBI) rats (A) was blocked by intrathecal 5, 7-dihydroxytryptamine (DHT) injection compared with vehicle (VEH) treated CNO/DREADD-injected TBI rats (B). In contrast, the analgesic response to DREADD-mediated activation of the PAG in CNO-treated, DREADD-injected, Uninjured rats was unaffected by intrathecal injection of the neurotoxin, 5, 7-dihydroxytryptamine (DHT; C). White arrow—prostaglandin E2 (PGE-2); intraplantar (i.pl.) [1 μg/50 μL]. Gray arrow—CNO; intraperitoneally (i.p.; 3 mg/kg). Four-way mixed repeated measures analysis of variance revealed a significant main effect of injury [F(1,48) = 7.93, p = 0.007, ƞ2 = 0.14], virus condition [F(1,48) = 188.93, p = 2.92 × 10⁻¹⁸, ƞ2 = 0.80], main effect of DHT [F(1,348) = 5.34, p = 0.025, ƞ2 = 0.10], time [F(5,240) = 74.63, p = 6.32 × 10⁻⁴⁷, ƞ2 = 0.61], injury by virus condition interaction [F(1,48) = 9.57, p = 0.003, ƞ2 = 0.17], injury by DHT [F(1.48) = 4.72, p = 0.035, ƞ2 = 0.09], virus condition by DHT [F(1,48) = 5.14, p = 0.028, ƞ2 = 0.10], injury by virus condition by DHT. Three-way interaction [F(1,48) = 7.87, p = 0.007, ƞ2 = 0.14], time by injury [F(5,240) = 4.07, p = 0.001, ƞ2 = 0.8], time by virus [condition F(5,240) = 74.69, p = 5.92 × 10⁻⁴⁷, ƞ2 = 0.61], time by injury by virus condition [F(5,240) = 5.83, p = 4.2 × 10⁻⁵, ƞ2 = 0.11]. Time by injury by DHT [F(5,240) = 2.33, p = 0.043, ƞ2 = 0.05], Time by injury by virus by DHT [F(5,240) = 2.35, p = 0.042, ƞ2 = 0.05]. No other main effects or interactions were significant. Post hoc analysis reveals that CNO anti-nociceptive effect is reversed by DHT in TBI animals but not in uninjured animals (panels B vs. C). Data are presented as mean ± standard error of the mean (n = 6-8).

Restoration of the DCN response in TBI rats pretreated with the selective serotonin reuptake inhibitor, escitalopram, is mediated via the 5-HT₇ receptor

We have demonstrated in a previous set of studies that systemic pretreatment with the SSRI escitalopram (ESC) for 3 days prior to DCN testing would restore the capsaicin-induced DCN behavioral response in TBI rats at 49 DPI.¹¹ With this in mind, we wanted to determine which serotonergic receptor was responsible for this restoration. A previous study showing restoration of the DCN response after a peripheral injury using an SSRI demonstrated that it was the 5-HT₇ receptor.³³ Therefore, we intrathecally injected ESC-pretreated TBI rats with the 5-HT₇ antagonist, SB-269970 (Fig. 7). We demonstrated that SB-269970 did block the capsaicin-induced DCN behavioral response in our ESC/TBI rats. It had no effect on the DCN response in ESC/Sham rats (Fig. 7). Three-way mixed repeated measures ANOVA revealed a significant main effect of injury condition (p = 7.96 × 10⁻¹⁰), SB-269970 (p = 7.96 × 10⁻⁸), time (p = 1.28 × 10⁻²⁵), injury by SB-269970 interaction (p = 1.5 × 10⁻⁴), time by injury interaction (p = 0.001), time by SB-269970 interaction (p = 9.7 × 10⁻⁵), time by injury by SB-269970 (p = 3.0 × 10⁻⁴). Post hoc analysis reveals that the restorative effect of escitalopram pretreatment is reversed by SB-269970 treatment (Fig. 7B). Data are presented as mean ± SEM (n = 8).

Intrathecal treatment with the 5-HT₇ antagonist, SB269970 does not block the analgesic response in CNO/DREADD-injected TBI rats

With the previous study in mind, we next wanted to see if the anti-nociceptive response in CNO/DREADD-injected, TBI rats was mediated via the 5-HT₇ receptor. Data from Figure 8B demonstrates that, unexpectedly, SB-269970 did not block the anti-nociceptive response in CNO/DREADD-injected TBI rats compared with VEH treated CNO/DREADD-injected TBI rats. Three-way mixed repeated measures ANOVA revealed a significant main effect of virus (p = 3.3 × 10⁻²⁸), time (p = 1.28 × 10⁻⁶⁶), time by virus condition interaction (p = 1.14 × 10⁻⁶⁵). No other main effects drug or interactions were significant. Post hoc analysis reveals that SB26 had no significant impact on the DREADD/CNO effect in TBI subjects (Fig. 8B). Three-way mixed repeated measures ANOVA revealed a significant main effect of virus (p = 3.3 × 10⁻²⁸], time (p = 1.28 × 10⁻⁶⁶), time by virus condition interaction (p = 1.14x × 10⁻⁶⁵). No other main effects drug or interactions were significant. Post hoc analysis reveals that SB-269970 had no significant impact on the DREADD/CNO effect in TBI subjects. Data are presented as mean ± SEM (n = 6-8).

FIG. 8.

The analgesic response to Designer Receptor Exclusively Activated by Designer Drug (DREADD)-mediated activation of the periaqueductal gray matter (PAG) in clozapine N-oxide (CNO)-treated, DREADD-injected traumatic brain injury (TBI) rats (A) was unaffected by the intrathecal treatment with 5-HT₇ receptor antagonist, SB-269970 (B). The analgesic response to DREADD-mediated activation of the PAG in CNO-treated, DREADD-injected, TBI rats was unaffected by the intrathecal treatment with 5-HT_1A receptor antagonist, WAY-100635 (C). White arrow—prostaglandin E2 (PGE-2); intraplantar (i.pl.) [1 μg/50 μL]. Gray arrow—CNO; intraperitoneally (i.p.; 3 mg/kg). Three-way mixed repeated measures analysis of variance revealed a significant main effect of virus [F(1,24) = 3997.11, p = 3.3 × 10⁻²⁸, ƞ2 = 0.99], time [F(7,168) = 139.66, p = 1.28 × 10⁻⁶⁶ ƞ2 = 0.85], time by virus condition interaction [F(7,168) = 135.44 p = 1.14 × 10⁻⁶⁵, ƞ2 = 0.85]. No other main effects drug or interactions were significant. Post hoc analysis reveals that SB26 had no significant impact on the DREADD/CNO effect in TBI subjects. (C) Three-way mixed repeated measures analysis of variance revealed a significant main effect of virus condition [F(1,26) = 1315.52, p = 8.51 × 10⁻²⁴, ƞ2 = 0.98], time [F(7,182) = 93.16, p = 9.28x10⁻⁵⁷ ƞ2 = 0.78], time by virus condition interaction [F(7,182) = 86.70 p = 1.43 × 10⁻⁵⁴ ƞ2 = 0.77]. No other main effects drug or interactions were significant. Post hoc analysis reveals that WAY-100635 had no significant impact on the DREADD/CNO effect in TBI subjects (C). Data are presented as mean ± standard error of the mean (n = 6-8).

Intrathecal treatment with the 5-HT_1A antagonist, WAY-100653 does not block the analgesic response in CNO/DREADD-injected TBI rats

To ascertain if 5-HT_1A receptors could be responsible for the anti-nociceptive response in the CNO/DREADD-injected TBI rats we used the 5-HT_1A antagonist, WAY-100653 (Fig. 8A). Data from Figure 8C reveal that intrathecal treatment with WAY-100653 failed to block the anti-nociceptive response in CNO/DREADD injected TBI rats compared with VEH treated CNO/DREADD-injected TBI rats. Three-way mixed repeated measures ANOVA revealed a significant main effect of virus condition (p = 8.51 × 10⁻²⁴), time (p = 9.28 × 10⁻⁵⁷), time by virus condition interaction (p = 1.43 × 10⁻⁵⁴). No other main effects drug or interactions were significant. Post hoc analysis reveals that WAY-100635 had no significant impact on the DREADD/CNO effect in TBI subjects (Fig. 8C). Data are presented as mean ± SEM (n = 6-8).

Intrathecal administration of the selective 5-HT_2A antagonist, ketanserin blocks the analgesic response in CNO/DREADD-injected TBI rats

To determine if the spinal 5-HT_2A receptor is involved in the anti-nociceptive response triggered in CNO/DREADD-injected TBI rats we used the 5-HT_2A antagonist, ketanserin (KET; Fig. 9A). Data from Figure 9B reveals that intrathecal treatment with KET did block the anti-nociceptive response in CNO/DREADD injected TBI rats compared with VEH treated CNO/DREADD-injected TBI rats. Three-way mixed repeated measures ANOVA revealed a significant main effect of virus condition (p = 4.04 × 10⁻¹⁷), ketanserin (p = 3.43 × 10⁻¹²), time (p = 2.49 × 10⁻⁴²), virus condition by ketanserin (p = 1.11 × 10⁻¹²), time by virus (p = 7.19 × 10⁻⁴¹), time by ketanserin (p = 1.50 × 10⁻²⁴), time by virus condition by ketanserin (p = 7.82 × 10⁻²⁸). Post hoc analysis reveals that ketanserin significantly reversed the DREADD/CNO effect in TBI subjects (Fig. 9B); *DREADD/TBI/CNO/KET vs. DREADD/TBI/CNO/VEH, p < 0.001) by Bonferroni's post hocs. Data are presented as mean ± SEM (n = 6-8).

FIG. 9.

The analgesic response to Designer Receptor Exclusively Activated by Designer Drug (DREADD)-mediated activation of the periaqueductal gray matter (PAG) in clozapine N-oxide (CNO)-treated, DREADD-injected, traumatic brain injury (TBI) rats (A) was blocked by the intrathecal treatment with 5-HT_2A receptor antagonist, ketanserin (B). White arrow—Prostaglandin E2 (PGE-2); intraplantar (i.pl.) [1 μg/50 μL]. Gray arrow— CNO; intraperitoneally (i.p.; 3 mg/kg). Three-way mixed repeated measures analysis of variance revealed a significant main effect of virus condition [F(1,26) = 385.83, p = 4.04 × 10⁻¹⁷ ƞ2 = 0.94], ketanserin [F(1,26) = 146.64, p = 3.43 × 10⁻¹², ƞ2 = 0.85], time [F(7,182) = 56.42, p = 2.49 × 10⁻⁴², ƞ2 = 0.69], virus condition by ketanserin [F(1,26) = 162.15, p = 1.11 × 10⁻¹², ƞ2 = 0.86], time by virus [F(7,182) = 53.39, p = 7.19 × 10⁻⁴¹, ƞ2 = 0.67], time by ketanserin [F(7,182) = 26.12, p = 1.50 × 10⁻²⁴, ƞ2 = 0.50], time by virus condition by ketanserin [F(7,182) = 30.75, p = 7.82 × 10⁻²⁸, ƞ2 = 0.54]. Post hoc analysis reveals that ketanserin significantly reversed the DREADD/CNO effect in TBI subjects (9B). *DREADD/TBI/CNO/Ketanserin (KET) vs. DREADD/TBI/CNO/Vehicle (VEH. p < 0.001) by Bonferroni's post hocs. Data are presented as mean ± standard error of the mean (n = 6-8).

Expression of spinal 5-HT_2A mRNA is not enhanced at 7 or 49 days post-TBI

Due to the effects of the lumbar intrathecal injections of ketanserin and having observed changes in spinal cord gene expression in response to TBI,^62,63 we use qPCR on rat lumber spinal cord samples to determine whether TBI enhanced 5-HT_2A expression. The results of these experiments failed to show an increase in expression of the 5-HT_2A receptor gene (Supplementary Fig. S1).

Intrathecal administration of the selective α-1 Adrenoceptor (AR) antagonist, prazosin does not block the analgesic response in CNO/DREADD-injected TBI rats

To determine if the spinal α-1 AR is involved in the anti-nociceptive response triggered in CNO/DREADD-injected TBI rats we used the α-1 AR antagonist, prazosin (Supplementary Fig. S2). Data from Supplementary Figure S2 reveals that intrathecal treatment with prazosin failed to block the anti-nociceptive response in CNO/DREADD injected TBI rats compared with VEH treated CNO/DREADD-injected TBI rats (Supplementary Fig S2). Post hoc analysis reveals that prazosin had no significant impact on the DREADD/CNO effect in TBI subjects (Supplementary Fig. S2). Data are presented as mean ± SEM (n = 6-8).

Discussion

Chronic pain is a consequence of TBI that frequently manifests as persistent headaches and pain to the trunk and extremities.^8,64,65 While research into the causes of chronic pain after TBI is still in its infancy, our research and that of others agree that an imbalance in the descending pain modulatory pathways is likely to be responsible.^{6,7,10,11,13,29,40} The disruption of pain modulatory pathways caused by diminution of descending inhibition and/or enhancement of descending facilitation would enable nociceptive signaling and promote the development of central sensitization, a key feature of patients with chronic pain.^{6,7,10–13,66–69}

The periaqueductal gray (PAG), a critical control center for descending pain modulation, receives inputs from several areas of the brain, including the hypothalamus, amygdala, and anterior cingulate cortex.^18,70 The PAG is internally delineated into four major longitudinal columns located dorsomedial, dorsolateral, lateral, and ventrolateral to the aqueduct. These delineations are based on anatomical connections and the functional representation of cardiovascular and behavioral functions they perform. The PAG was selected as our location for the DREADD virus because it and the neighboring dorsal raphe nucleus are known originators of potent analgesia and opioid-induced anti-nociception.^{24,25,71–75} Activation of these structures produces active and passive coping responses to the environment, permitting use of adaptive anti-nociceptive or locomotor strategies to lessen the effect of pain.^76,77

Outputs from the PAG descend indirectly to medullary or spinal dorsal horns via the rostral ventromedial medulla (RVM).^14-16 Both descending serotonergic pain facilitation and inhibition can result from this pathway, which depends on what serotonergic receptors are activated within the dorsal horn (i.e., activation of 5-HT3 facilitates pain, whereas activation of 5-HT_1A, 5-HT_2A, or 5-HT₇ will result in pain inhibition).^17–20 The PAG is also rich in opioid receptors and research has shown that injection of opioids into the PAG will result in antinociception. Further, antinociception resulting from the electrical stimulation of the PAG has been shown to be dependent on an intact opioid receptor mediated pathway in both preclinical and human patients. Both the PAG and the RVM communicate with noradrenergic sites that are also important for pain modulation, including A5, the locus coeruleus (A6), and A7. These noradrenergic nuclei are the major sources of direct noradrenergic projections to the spinal cord.⁷⁸ The noradrenaline released by these descending fibers binds to α-2-adrenoceptors within the dorsal horn to inhibit nociceptive signaling.⁷⁸

In this study we assessed the anti-nociceptive efficacy and mechanisms of the PAG in uninjured and TBI injured rats using a DREADD virus technology. We found that activation of the DREADD-injected PAG with CNO could generate an anti-nociceptive response to PGE-2-induced sensitization of the hindlimb in uninjured rats that could be blocked with the opioid receptor antagonist, naloxone. We then confirmed that the pattern of nociceptive changes observed in the 7 weeks following TBI was unchanged in the DREADD-injected TBI rats suggesting that this tool was unlikely to alter the normally observed post-TBI pain physiology (Fig. 2). Having established the viability of this DREADD tool, we assessed the ability of DREADD-mediated PAG activation to block nociceptive sensitization during the initial post-TBI pain phase. We found that CNO/DREADD-mediated activation of the PAG in TBI rats did provide a transient (240 min) and significant increase in paw withdrawal threshold at both 7 and 21 days post-TBI (Fig. 3).

The expression of DREADD constructs is sufficiently stable to study long-term neurophysiological changes after TBI.⁷⁹ Even 49 days after injury, CNO administration caused an anti-nociceptive response after sensitizing hindpaws with PGE-2 in PAG DREADD-expressing TBI rats (Fig. 4). These effects are likely due to the activated DREADD as this effect was not observed in DREADD-injected TBI rats given the vehicle. Moreover, an anti-nociceptive response was not observed in control virus-injected TBI rats regardless of the presence of CNO. These are key data for our experiments and others considering the use of similar techniques as there have been some reports that CNO alone can influence an animal's behavior.^80-82 However, with our balanced experimental design, we were largely able to exclude the possibilities that CNO alone or DREADD-expression itself interfered with the behavioral paradigm.

Unexpectedly, systemic treatment with naloxone failed to block the anti-nociceptive effect of the CNO/DREADD-mediated activation of the PAG to hindpaw hypersensitivity in TBI rats at 49 DPI, thus indicating that the CNO/DREADD-mediated anti-nociceptive response was no longer dependent on an opioid-receptor mediated pathway after TBI. Behavioral anti-nociception produced following PAG stimulation has also been shown to be mediated in part by the serotonin.^83-85 To investigate the role of serotonin in the anti-nociceptive response mounted in CNO/DREADD-injected TBI rats we selectively depleted the serotonergic fibers in the lumbar spinal cord with 5, 7-dihydroxytryptamine (DHT). This revealed that the anti-nociceptive response in CNO/DREADD-injected TBI rats was dependent on a serotonin driven mechanism. However, eliminating spinal serotonergic fibers did not block the anti-nociceptive response stimulated in the CNO/DREADD-injected uninjured rats.

To identify the serotonergic receptor responsible for the anti-nociceptive response in CNO/DREADD-injected TBI rats we first investigated the role of 5-HT₇. This decision was based on the finding that the restoration of endogenous nociceptive modulation in TBI rats following pre-treatment with escitalopram, exemplified by testing for the presence of an intact DCN response, was dependent on intact 5-HT₇ receptor signaling. Despite this finding the 5-HT₇ receptor antagonist, SB-269970, failed to block the anti-nociceptive response in CNO/DREADD-injected TBI rats. A second serotonin receptor, 5-HT_1A, is commonly associated with antinociception in several pain assays in the rat.^86-88 However, the anti-nociceptive response in the CNO/DREADD-injected TBI rats still remained following treatment with the 5-HT_1A antagonist, WAY-100635.

Lastly, activation of spinal 5-HT_2A receptors have been shown to suppress allodynia in both a rat model of spinal nerve ligation and to inflammatory stimuli.^47,89 When DREADD-expressing TBI rats were spinally injected with the 5-HT_2A receptor antagonist ketanserin, the CNO-mediated anti-nociceptive response was significantly reduced. Therefore prior to TBI, the anti-nociceptive response in CNO/DREADD-injected rats was an opioidergic sensitive mechanism. However, following a TBI the mechanism responsible for antinociceptive response to PGE-2 hindlimb hypersensitivity, switched to a 5-HT_2A receptor-mediated pathway. We have similarly observed a change in neural circuitry after brain injury using DREADD constructs specific to neurons of the locus coeruleus.⁴³ In those studies, we demonstrated that following TBI, animals lose descending α₂ inhibition of nociception but gain a pathway that involves α₁ inhibition of nociception. Tests using the α₁ AR antagonist, prazosin, in TBI rats injected with the DREADD virus into the PAG revealed that prazosin failed to block the anti-nociceptive response following CNO treatment. The findings of fundamental signaling changes in two of the brainstem's most powerful descending nociceptive modulation systems serve as examples of the plasticity TBI can induce. It is acknowledged, however, that 5-HT_2A agonism is responsible for the hallucinogenic properties of some drugs, and that this may be a factor in 5-HT_2A -directed drug development.

Designer Receptor Exclusively Activated by Designer Drugs or DREADDS are an adeno-associated virus (AAV) that include specific promoters to limit their expression to particular cell populations.^48–51 The AAV used in the current study contains the human synapsin promoter, hSyn,⁵⁴ to selectively express the excitatory hM3Dq DREADD receptor⁵² in neurons of the injected area, the PAG. Besides our aforementioned study using DREADDs specific to the neurons of the LC,⁴³ these tools have seldom been used to define or treat the consequences of TBI.⁹⁰

There are limitations to this study. For example, there are concerns that CNO and its metabolic products (clozapine and N-desmethylclozapine) could have off-target effects on behavior.^80-82 However, we were cognizant of including all the appropriate control groups (i.e., CNO/Control-injected rats) and we observed no behavioral effects in rats treated with CNO in the absence of a DREADD. It is also possible that expression of the DREADD virus may have spread to other neuronal cell populations outside of the PAG. However, mCherry-positive cells and their axons were exclusively found within the PAG. Another limitation is the exclusive use of male rats in this study. We are currently investigating the manifestation of nociception in female rats after a TBI and early results suggest that females do exhibit the same two-phases of nociception as observed in males (unpublished observations). However, future studies are required to determine what mechanisms are involved in the manifestation of nociception in female TBI rats. We acknowledge that a robust pre-clinical evaluation of the strategy of using 5-HT_2A agonists as analgesics in the setting of TBI will require the use of well-validated pain models. Models validated for the study of headache, back and joint pain will be particularly important. Taken as a whole, the results strongly suggest that neuromodulation of descending projections emanating from the PAG reflect a viable target for treating post-TBI pain. Future studies should include investigations into the mechanisms initiated by trauma that led to these signaling changes and use them as a foundation for designing interventions to control them and improve TBI outcomes.

Transparency, Rigor, and Reproducibility Summary

All studies were approved by the Veterans Affairs Palo Alto Health Care System Institutional Animal Care and Use Committee (Palo Alto, CA, USA) and followed the animal guidelines of the National Institutes of Health Guide for the Care and Use of Laboratory animals (NIH Publications 8th edition, 2011).⁴⁴ A total of 271 Sprague Dawley rats were randomly divided into different groups. All data are shown as the mean ± SEM. An independent statistician (A.R.F.) tested statistical assumptions, performed missing values analysis, descriptive statistical tests, followed by inferential tests. Data met normality, homogeneity of variance, and independence assumptions sufficiently for application of general linear models (GLM). Group comparisons were therefore performed according to a priori balanced experimental designs as two-way, three-way, or four-way mixed ANOVAs, modeling between- and within-subject factors as appropriate using GLM. Significant effects were followed by interaction plots and Bonferroni's post hocs. All statistical analyses were performed using IBM SPSS software (v.27, IBM Corp.) with base, regression, missing values, and advanced statistics modules. Statistical code is freely available upon request. Graphs were created using Prism 9.1.0 (GraphPad Software). Data are presented as mean values ± standard error of the mean (SEM). All experimental sample sizes (n) were selected by a priori power calculations based on historical data from our laboratory. The figure legends report precise F values, degrees of freedom, p values, effect sizes, and observed power for all significant effects as well as the n's completed for each group. Experimenters were blinded to the identity of treatments and experimental conditions. All studies were designed to minimize the number of rats required. Before testing, all rats were also transferred into the testing room to acclimate to the environment for 60 min to ensure consistent activity. Mechanical withdrawal thresholds were measured using a modification of the up-down method and von Frey filaments as described previously.⁵⁰ Our investigators were proficiency trained in advance to use the von Frey apparatus.

Authors' Contributions

Karen-Amanda Irvine: conceptualization, methodology, investigation, data curation, simple analysis, original draft, review, and editing. Xiao-You Shi: PCR data curation and analysis. Adam R Ferguson: formal analysis, review, and editing. J. David Clark: conceptualization, methodology, original draft, review, and editing.

Author Disclosure Statement

No competing financial interests exist.